Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method for operating a graphics processing unit (GPU), comprising: generating, using the GPU, a result retained within GPU-accessible memory based at least in part on receiving sensor input data from a sensor, the generating occurring without providing the sensor input data from the GPU to a central processing unit (CPU); rendering, using the GPU, a display based at least in part on accessing the result from the GPU-accessible memory; and providing, using the GPU, a canonical memory location within the GPU-accessible memory storing the result to an application or an operating system component executing on the CPU after rendering the display.
A method for operating a GPU involves the GPU directly processing sensor data (like camera or touch input) to create a result, such as a gesture recognition. This result is stored in memory accessible by the GPU. The GPU then uses this result to render a display. After the display is rendered, the GPU provides the memory location where the result is stored to an application or the operating system running on the CPU. This avoids sending the sensor data to the CPU for processing, reducing latency.
2. The method of claim 1 , wherein the GPU does not access the CPU during the generating the result.
Building upon the method where a GPU directly processes sensor data and renders a display based on the result, the GPU does not need to access the CPU at all during the processing of the sensor data to generate the result. All processing happens entirely on the GPU.
3. The method of claim 1 , wherein the sensor input data comprises touch sensor input data; and wherein the sensor is a touch sensor.
In the method where a GPU directly processes sensor data and renders a display, the sensor data specifically comprises touch sensor input from a touch sensor, like a touchscreen on a phone or tablet. The GPU processes this touch input directly to generate a result used for display.
4. The method of claim 1 , further comprising generating, using the GPU, the result with a touch-sensor processing GPU code component.
In the method where a GPU directly processes sensor data and renders a display, a specific component of GPU code dedicated to touch-sensor processing is used to generate the result. This component handles the specific logic for interpreting touch input.
5. The method of claim 1 , further comprising rendering, using the GPU, the display with a graphics rendering GPU code component.
In the method where a GPU directly processes sensor data and renders a display, a dedicated graphics rendering GPU code component handles the rendering of the display based on the result generated from the sensor data.
6. The method of claim 1 , wherein the result is not provided by the CPU to the GPU.
In the method where a GPU directly processes sensor data and renders a display, the result generated by the GPU from the sensor data is not provided *by* the CPU to the GPU at any point. The GPU is entirely responsible for generating the result from the sensor input.
7. The method of claim 1 , wherein the sensor input data corresponds to at least one of touch input, camera input, infrared input, audio input, video input, stylus input, or depth input.
In the method where a GPU directly processes sensor data and renders a display, the sensor data can be touch input, camera input, infrared input, audio input, video input, stylus input, or depth input. The GPU directly processes these various types of sensor data.
8. The method of claim 1 , wherein the result corresponds to a gesture derived from the input sensor data.
In the method where a GPU directly processes sensor data and renders a display, the result derived from the sensor data is specifically a gesture. For instance, touch input might be interpreted as a swipe, pinch, or tap gesture.
9. The method of claim 8 , further comprising: invoking an output process to render a user interface on the display based upon the gesture.
Building upon the method of directly processing sensor data to recognize a gesture, an output process is invoked on the GPU to render a user interface element on the display that corresponds to the recognized gesture. For example, a swipe gesture could trigger a screen transition.
10. The method of claim 1 , further comprising: storing the result within the canonical memory location of the GPU-accessible memory.
Building upon the method where a GPU directly processes sensor data and renders a display, the result is stored within a pre-defined, known (canonical) memory location inside the GPU's accessible memory. This allows other processes to easily find the result.
11. The method of claim 10 , further comprising: providing the canonical memory location asynchronously to rendering the display.
Building upon the method of storing the result in a known memory location, the location of this memory is provided to the CPU asynchronously, meaning it is not necessarily provided at the same time as the display is rendered. This allows the CPU to read the data later without blocking the rendering process.
12. The method of claim 1 , further comprising: providing an input process with read/write access to the sensor input data; and providing an output process with at least read access to at least one of the sensor input data or the result.
In the method where a GPU directly processes sensor data and renders a display, an input process is given both read and write access to the raw sensor data. An output process is then given at least read access to either the original sensor data or the processed result. This allows both processes to operate independently on the data.
13. The method of claim 1 , comprising: scheduling, by an operating system scheduler, the generating the result or rendering the display for execution using the GPU.
In the method where a GPU directly processes sensor data and renders a display, the operating system scheduler is responsible for scheduling the tasks of generating the result from the sensor data and rendering the display for execution on the GPU. The OS manages when these processes are run.
14. A system for operating a graphics processing unit (GPU), comprising: a GPU scheduler configured to: generate a result retained within GPU-accessible memory based at least in part on receiving sensor input data from a sensor, the generating occurring without providing the sensor input data from the GPU to a central processing unit (CPU); render a display based at least in part on accessing the result from the GPU-accessible memory; and provide a canonical memory location within the GPU-accessible memory storing the result to an application or an operating system component executing on the CPU after rendering the display.
A system for operating a GPU includes a GPU scheduler. The scheduler directly processes sensor data to generate a result, stores that result in GPU memory, and renders a display based on that result. The sensor data isn't sent to the CPU for this processing. After rendering, the scheduler provides the memory location of the result to an application or OS component running on the CPU.
15. The system of claim 14 , wherein the scheduler is further configured to: use the result before providing the result to the CPU.
In the GPU system using a scheduler to process sensor data and render a display, the scheduler is configured to use the result for rendering the display before giving access of the result data to the CPU to further reduce the display latency.
16. The system of claim 14 , further comprising: a data access interface configured to: provide the canonical memory location to the CPU asynchronous to rendering the display.
In the GPU system that uses a scheduler, a data access interface provides the CPU with the memory location of the result. The access to this location is asynchronous, meaning it doesn't happen at the same time as the rendering of the display. The CPU can access this data later without blocking the rendering process.
Unknown
October 3, 2017
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.